v0.5.0

Highlights

• Support on-premise deployment.
• Provide RAG services using PGVector and large languange models (data source type: RagApi), compatible with standard OpenAI protocols, and support operations using Cherry Studio.
• Support data migration from files (data source type: SshFile, S3File, OssFile) to PostgreSQL and text embedding using embedding models. Combined with RagApi, RAG applications based on private data can be built.
• Support PostgreSQL text embedding using embedding models and storing embeddings in PGVector.
• Support MySQL text embedding using embedding models and storing embeddings in PGVector.
• Support adding OpenAI, Alibaba Cloud DashScope, DeepSeek, HuggingFace, Cohere, LocalAI and other LLM type data sources for text embedding and RagApi services.

New Connections

• Support Dameng -> MySQL/StarRocks/Doris schema migration, full data migration, incremental data sync, data verification and correction, DDL sync (insert/delete columns).
• Support Dameng -> ClickHouse schema migration, full data migration, incremental data sync, data verification, DDL sync (insert/delete columns).
• Support Dameng -> Kafka full data migration and incremental data sync.
• Support full data migration from file data sources (SshFile, S3File, OssFile) to PostgreSQL.
• Support full data migration from file data sources (SshFile, S3File, OssFile) to MySQL.
• Support Greenplum -> Hana schema migration and full data migration. It also supports scheduled full data migration and filtering conditions (in-program filtering/predicate pushdown).
• Support Greenplum -> Doris schema migration, full data migration, and data verification and correction.

New Features

• Support POSIX file system files obtained over SSH (SshFile), AWS S3 files (S3File), Alibaba Cloud OSS files (OssFile) data sources.
• Support full data migration from text files (txt, markdown, json, .java, .c, etc.), Excel, CSV to MySQL or PostgreSQL.
• Support MariaDB v11.8 (actually BladePipe is compatible with all 11.x versions).
• Support Aliyun RocketMQ v5.0.
• Add a static mode (multi-offset) for SQL Server CDC table, and DataJobs share a fixed CDC table name (db_schema_table_cc_static).
• Support configuring target RDB parallel write strategy in user preference setting (parameter: increParallelApplyStrategy).
• Support custom code for an Iceberg instance.
• Support filtering Redis keys using regular expression (new parameter: keyRegex).
• Allow setting the target primary key in a source MongoDB instance.
• At the table selection step of DataJob creation, users can search and manually select tables to meet the needs of limited database authorization and selecting a few tables from hundreds of thousands of tables for data replication.
• Allow to configure whether to obtain unique key information when creating a DataJob in user preference setting.
• Automatically select UPDATE & DELETE or making the table selectable for a table with unique keys but no primary keys when creating a DataJob, and automatically set the first unique key field as the target primary key on the column selection page.
• Support running BladePipe using JDK 17, and overall operating performace is optimized.
• Allow clearing up Worker logs.

Improvements

• Improve Redis offset, add count auxiliary index and allow dbIndex multiple offsets.
• Improve Alibaba Cloud RocketMQ connection test, which can detect whether the RocketMQ instance is alive.
• The offsets of a source Alibaba Cloud RocketMQ instance can be reset by timestamp.
• Optimize SQL Server metadata acquisition, showing the acquisition speed.
• Support multiple TiSession requests for a TiDB source to address DataJob Region processing bottleneck (parameter: maxPoolSize).
• Optimize large key sync in full data migration from Redis (new parameters: listValueShardSize, hashValueShardSize, setValueShardSize, zsetValueShardSize).
• Improve Redis offset, add count auxiliary indicators, and multiple offsets for dbIndex.
• Allow to manually create date type format when there's no format of data writing to ElasticSearch.
• Improve time zone conversion in data sync and time zone data verification and correction for a target ElasticSearch instance.
• Delete the default Worker in the first installation of BladePipe.
• Improve the display of data source list, easier to identify.

Bug Fixes

• Fix the problem that the source Redis master node is unstable due to multiple restarts of a Redis -> Redis DataJob (new parameter: bindSocketPort).
• Fix the error that the Socket connection isn't closed when the Redis -> Redis DataJob is stopped.
• Fix the exception occurred in the incremental data sync when the Kafka Scram login mode and the raw message format are selected.
• Fix the problem that the DDL cannot be captured due to the uppercase table name on the TiDB source.
• Fix the problem that the number of tables cannot be displayed when modifying subscriptions.
• Fix the error of conversion of largeint type data by StarRocks DYNAMIC scanMode.
• Fix the null pointer error during PolarDbMySQL full data migration.
• Fix the exception occurred in subscription modification due to long Dbs in Hana, PostgreSQL, Dameng, TDengine, SQL Server.
• Fix the illegal groupid exception occurred in subscription modification of a source Alibaba Cloud RocketMQ instance.
• Fix the problem of configuration merge failure in subscription modification of a source Alibaba Cloud RocketMQ instance.